An adaptive transform coding algorithm for a still image utilizes a quadtree based variable block size discrete cosine transform to achieve a better tradeoff between bit rate and image quality. The choice of appropriate block size is made according to a mean based decision rule which can discriminate various image contents for better visual quality.
|
16. A method for processing a signal comprising the steps of
providing a signal comprising a two-dimensional array of pixels which is representative of a still image, dividing said signal comprising a two-dimensional array of pixels into blocks of pixels of size M×M, each of said M×M blocks of pixels having a content, decomposing each of said M×M blocks into one or more sub-blocks, comparing adjacent sub-blocks and merging said adjacent sub-blocks when the edges of said adjacent sub-blocks are not in high contrast, applying an orthogonal transform to each of said sub-blocks to form transform coefficients, compressing the transform coefficients formed for each of said sub-blocks to form compressed transform coefficients, utilizing an encoder to form a bit stream corresponding to the compressed transform coefficients, and storing the bit stream in a memory.
14. A method for processing a signal comprising the steps of
providing a signal comprising a two-dimensional array of pixels which is representative of a still image, dividing said signal comprising a two-dimensional array of pixels into blocks of pixels of size M×M, each of said M×M blocks of pixels having a content, selectively decomposing each of said M×M blocks into one or more sub-blocks, comparing adjacent sub-blocks and merging said adjacent sub-blocks when the edges of said adjacent sub-blocks are not in high-contrast, applying the discrete cosine transform to each of said sub-blocks to form transform coefficients, compressing the transform coefficients formed for each of said sub-blocks to form compressed transform coefficients, utilizing an encoder to form a bit stream corresponding to the compressed transform coefficients, and passing the bit stream to a channel coder for transmission via a channel to a receiver at a remote location.
13. A method for processing a signal corresponding to a still image comprising the steps of
providing a signal comprising of two dimensional array of pixels which is representative of a still image, dividing said signal comprising said two-dimensional array of pixels into blocks of pixels of varying sizes such that the size of said blocks is determined by finding boundaries where the abutting edges of said varying sized blocks are high contrast edges, applying the orthogonal transform to each of said blocks of pixels to form a two-dimensional array of transform coefficients for each of said blocks of pixels, compressing the transform coefficients obtained for each block of pixels to form a two-dimensional array of compressed transform coefficients for each block of pixels, scanning the two-dimensional array of compressed transform coefficients to form a signal comprising one-dimensional sequence of transform coefficients for each block of pixels, utilizing an encoder to encode the signal comprising a one-dimensional sequence of transform coefficients for each block to form a bit stream representative of a still image, and transmitting said bit stream via a channel to a receiver at a remote location.
1. A method for processing a signal corresponding to a still image so that said still image can be transmitted to a remote location via a channel, said method comprising the steps of
providing a signal comprising a two-dimensional array of pixels which is representative of a still image, dividing said signal into blocks of pixels of size M×M, each of said M×M blocks of pixels having a content, splitting each of said M×M blocks to produce a hierarchy of sub-blocks including the levels designated i-0,1,2, . . . ,K-1, so that at each level of the hierarchy, there are 4i sub-blocks of size m/2i ×M/2i, i=0,1, . . . ,K-1, where K is an integer, starting with the K-1 level in said hierarchy and proceeding to successive levels in said hierarchy, merging adjacent of said sub-blocks when the differences in the mean pixel value of said sub-blocks is below a predetermined threshold, applying an orthogonal transform individually to each of said sub-blocks comprising said M×M blocks of pixels to form transform coefficients for each of said sub-blocks, utilizing an encoder to convert the transform coefficients for each sub-block into a signal in the form of a stream of bits, and transmitting said signal in the form of a stream of bits for each of said sub-blocks via a channel to a receiver at a remote location.
9. A method for processing a signal corresponding to a still image so that said still image can be transmitted to a remote location via a channel, said method comprising the steps of
providing a signal comprising a two-dimensional array of pixels which is representative of a still image, dividing said signal into blocks of pixels of size M×M, each of said M×M blocks of pixels having a content, depending on the content of each of said M×M blocks of pixels, selectively decomposing each of said M×M blocks of pixels into one or more sub-blocks of pixels, each of said sub-blocks being of a size given by m/2i ×M/2i, where i=0,1,2, . . . ,K-1 and where K is an integer, said decomposing step comprising splitting each of said M×M blocks to produce a hierarchy of sub-blocks including the levels designated i=0,1, . . . K-1, so that at each level of the hierarchy, there are 4i sub-blocks of size m/2i ×M/2i, i=0,1, . . . K-1, and starting at the level of K-1 in said hierarchy and proceeding to successive levels in said hierarchy, testing each set of four childless brother/sister sub-blocks and merging each set of four childless brother/sister sub-blocks in which the difference of mean pixel value of any two sub-blocks is smaller than a predetermined value until a level in the hierarchy is reached wherein there are no sets of childless brother/sister sub-blocks to be tested, applying an orthogonal transform individually to each of said sub-blocks comprising said M×M blocks of pixels to form transform coefficients for each of said sub-blocks, utilizing an encoder to convert the transform coefficients for each sub-block into a signal in the form of a stream of bits, and transmitting said signal in the form of a stream of bits for each of said sub-blocks via a channel to a receiver at a remote location.
10. A method for processing a signal corresponding to a still image so that said still image can be transmitted to a remote location via a channel, said method comprising the steps of
providing a signal comprising a two-dimensional array of pixels which is representative of a still image, dividing said signal into blocks of pixels of size M×M, each of said M×M blocks of pixels having a content, depending on the content of each of said M×M blocks of pixels, selectively decomposing each of said M×M blocks of pixels into one or more sub-blocks of pixels, each of said sub-blocks being of a size given by m/2i ×M/2i, where i=0,1,2, . . . , K-1 and where K is an integer, said decomposing step comprising splitting each of said M×M blocks to produce a hierarchy of sub-blocks including the levels designated i=0,1, . . . K-1, so that at each level of the hierarchy, there are 4i sub-blocks of size m/2i ×M/2i, i=0,1, . . . K-1, and starting at the level of K-1 in said hierarchy and proceeding to successive levels in said hierarchy, testing each set of four childless brother/sister sub-blocks and merging each set of four childless brother/sister sub-blocks in which the difference of mean pixel value of any two sub-blocks is smaller than a predetermined value, chosen individually for each level in said hierarchy, until a level in the hierarchy is reached wherein there are no sets of childless brother/sister sub-blocks to be tested, applying an orthogonal transform individually to each of said sub-blocks comprising said M×M blocks of pixels to form transform coefficients for each of said sub-blocks, utilizing an encoder to convert the transform coefficients for each sub-block into a signal in the form of a stream of bits, and transmitting said signal in the form of a stream of bits for each of said sub-blocks via a channel to a receiver at a remote location.
11. A method for processing a signal corresponding to a still image comprising the steps of
providing a signal comprising a two dimensional array of pixels which is representative of a still image, dividing said signal comprising said two-dimensional array of pixels into blocks of pixels of varying sizes such that smooth areas of said image are smaller sized blocks and busy areas of said image are larger sized blocks, said dividing step comprises dividing said image into M×M blocks of pixels, splitting said M×M blocks to produce a hierarchy of sub-blocks including the levels designated 1=0,1, . . . K-1, where K is an integer, so that at each level of the hierarchy there are 4i sub-blocks of size m/2i ×M/2i, and starting with the K-1 in the hierarchy and proceeding to successive levels in the hierarchy, testing each set of four childless brother/sister sub-blocks and merging each set of four childless brother/sister sub-blocks and merging each set of four childless brother/sister sub-blocks in which the difference of mean pixel value of any two sub-blocks is smaller than a predetermined value until a level in the hierarchy is reached in which there are no sets of four childless brother/sister sub-blocks to be tested, applying the orthogonal transform to each of said blocks of pixels to form a two-dimensional array of transform coefficients for each of said blocks of pixels, compressing the transform coefficients obtained for each block of pixels to form a two-dimensional array of compressed transform coefficients for each of said blocks of pixels, scanning the two-dimensional array of compressed transform coefficients to form a signal comprising a one-dimensional sequence of transform coefficients for each block of pixels, utilizing an encoder to encode the signal comprising a one-dimensional sequence of transform coefficients for each block to form a bit stream representative of a still image, and transmitting said bit stream via a channel to a receiver at a remote location.
12. A method for processing a signal corresponding to a still image comprising the steps of
providing a signal comprising a two dimensional array of pixels which is representative of a still image, dividing said signal comprising said two-dimensional array of pixels into blocks of pixels of varying sizes such that smooth areas of said image are smaller sized blocks and busy areas of said image are larger sized blocks, said dividing step comprising dividing said image into M×M blocks of pixels, splitting said M×M blocks to produce a hierarchy of sub-blocks including the levels designated 1=0,1, . . . K-1, where K is an integer, so that at each level of the hierarchy there are 4i sub-blocks of size m/2i ×M/2i, and starting with the K-1 in the hierarchy and proceeding to successive levels in the hierarchy, testing each set of four childless brother/sister sub-blocks and merging each set of four childless brother/sister sub-blocks and merging each set of four childless brother/sister sub-blocks in which the difference of mean pixel value of any two sub-blocks is smaller than a predetermined value, determined individually for each level in the hierarchy, until a level in the hierarchy is reached in which there are no sets of four childless brother/sister sub-blocks to be tested, applying the orthogonal transform to each of said blocks of pixels to form a two-dimensional array of transform coefficients for each of said blocks of pixels, compressing the transform coefficients obtained for each block of pixels to form a two-dimensional array of compressed transform coefficients for each of said blocks of pixels, scanning the two-dimensional array of compressed transform coefficients to form a signal comprising a one-dimensional sequence of transform coefficients for each block of pixels, utilizing an encoder to encode the signal comprising a one-dimensional sequence of transform coefficients for each block to form a bit stream representative of a still image, and transmitting said bit stream via a channel to a receiver at a remote location.
4. The method of
5. The method of
6. The method of
7. The method of
8. The method of
15. The method of
|
|||||||||||||||||||||||||||||
This application is a continuation of application Ser. No. 07/389,420, filed Aug. 7, 1989, abandoned.
A patent application entitled "Adaptive Transform Coding of Still Images", filed for C. T. Chen and D. J. LeGall on Jul. 19, 1989 and bearing Ser. No. 07/381,860 is assigned to the assignee hereof and contains subject matter related to the subject matter of the present application.
The present invention relates to a method for coding still images. More particularly, the present invention relates to a transform coding algorithm for an image in which the image is divided into pixel blocks of different size according to image contents and in which an orthogonal transform such as a discrete cosine transform is applied to the pixel blocks.
Transform coding has proven to be an efficient means of coding a still image signal for transmission to a remote receiver.
In accordance with a basic transform image coding algorithm, a still image is divided into blocks of pixels. For example, each block of pixels may be a 16×16 block. Each block of pixels undergoes a two-dimensional orthogonal transform to produce a two-dimensional array of transform coefficients. For still image coding applications, the discrete cosine transform has emerged as the best orthogonal transform from the point of view of image compression and ease of implementation.
After a block of pixels undergoes a transform such as a discrete cosine transform, the resulting transform coefficients are subject to compression by thresholding and quantization. Thresholding involves setting all coefficients whose magnitude is smaller than a threshold value T equal to zero. In some thresholding processes, when a coefficient has a magnitude greater than T, T is subtracted if the coefficient is positive and T is added if the coefficient is negative. The DC transform coefficient may be excluded from the thresholding process. Quantization involves scaling a coefficient by a step size Q and rounding off to the nearest integer.
After compression, the compressed transform coefficients are then coded, for example, using a Huffman code, to form a bit stream for transmission to a remote receiver. At the receiver, an inverse process occurs to reconstruct the original block pixels. Thus, the received bit stream is decoded and the received quantized transform coefficients are expanded by the quantization step size. An inverse transform such as an inverse discrete cosine transform is then applied to the transform coefficients to reconstruct the block of pixels. In this manner, an image is transmitted block by block from a transmitter to a receiver.
In the conventional transform coding algorithm described above, the parameters of importance are the block size, the threshold value T, and the quantization step size Q. The conventional transform coding algorithm uses fixed parameters for all blocks from an image, i.e., the block size, threshold value T and quantization step size Q remain constant for all blocks within an image and from image to image.
The drawback of using fixed parameters is that the bit rate and image quality cannot achieve their best trade-off for a wide variety of images. Higher image quality is obtained at the price of higher bit rate for those busy image areas where the same visual quality may be achieved at a lower bit rate with another set of parameters.
To yield a better tradeoff between the bit rate and image quality, the parameters used in the transform coding algorithm have to be adaptive to the image contents. This motivates an adaptive transform coding approach for still images, where higher image quality can be obtained at a lower average bit rate (see e.g., W. H. Chen et al, "Scene Adaptive Coder", IEEE Trans. Commun., Vol COM-32, pp 225-232, March 1984; N. Texier et al "Hybrid Transform Video Coding Scheme for Low Bit Rate: An Improvement of the Reference Algorithm", Proc. Picture Coding Symp. pp. 8.4-1-2, Turin, Italy, September 1988).
In the above-identified patent application, an adaptive transform coding algorithm is disclosed which utilizes an adaptive threshold value T and an adaptive quantization step size Q along with a fixed block size. In particular, the compression parameters Q and T are selected individually for each block of pixels. In general, smaller (larger) compression parameters result in higher (lower) bit rates and image quality. Thus, large compression parameters are used for busy blocks where the human visual system is not very sensitive to degradation and small compression parameters are used for smooth blocks where the human visual system to degradation.
Illustratively, a measure of the busyness of a block is the magnitude of the (K+1)th most significant transform coefficient, where K is an integer which remains the same for all blocks in an image or in a plurality of images. The magnitude of the (K+1)th most significant transform coefficient is used to determine the threshold value T and quantization step size Q individually for each block of pixels. The magnitude of the (K+1)th transform coefficient is large for busy blocks where the human visual system is not sensitive to degradation and results in large compression parameters and lower bit rates. On the other hand, for smooth blocks, the transform coefficients fall off rapidly, and the magnitude of the (K+1)th transform coefficient is small resulting in small compression parameters and higher bit rates. In comparison to the conventional transform coding algorithm, the adaptive transform coding algorithm described above provides more consistent image quality and lower average bit rate for a wide variety of images.
It is an object of the present invention to provide an alternative adaptive transform coding algorithm wherein adaptivity is achieved by varying the pixel block size according to image contents.
In particular, a larger pixel block size advantageously results in better compression because more pixels are provided for decorrelation in the transform process. However, an inherent disadvantage of a large block size is that the quality of the resulting reconstructed image is not consistent amongst various images or from block to block within a single image. For example, blocking and ringing artifacts are more visible in a reconstructed image where a relatively larger pixel block size is used if the image contains high contrast edges. On the other hand, small pixel block size enables less compression but better quality of the reconstructed image.
Therefore, it is a further object of the present invention to provide an adaptive transform coding algorithm which takes advantage of the compression capability of large block size and the high visual image quality of small block size by choosing the appropriate block size according to the image contents so as to achieve a better tradeoff between bit rate and image quality.
The present invention is an adaptive transform coding algorithm wherein adaptivity is achieved via a variable block size determined according to image contents.
An image is first divided into relatively large blocks of pixels of size M×M. Let N×N be the size of the smallest block of pixels of interest in the inventive algorithm, where M=N×2K-1 and where K is a positive integer. Each M×M block is decomposed into 4K-1 blocks of size N×N by successively splitting each block into four equally sized sub-blocks until the smallest sub-block of size N×N is reached.
The successive splitting of blocks results in a hierarchy of sub-blocks with K levels, which levels are designated i=0,1,2 . . . K-1. At each level in the hierarchy, the sub-block size is M/2i ×M/2i. All sub-blocks at the same level of the hierarchy are in the same "generation" and sub-blocks derived from the same parent sub-block are said to be brother/sister sub-blocks.
Illustratively, a three-level hierarchy of sub-blocks (i.e. K=3) is utilized. For example, if M=16, the i=0 level of the hierarchy comprises one 16×16 block. The i=1 level of the hierarchy comprises four 8×8 brother/sister sub-blocks arrived at by splitting the 16×16 block into four sub-blocks. The i=K-1=2 level of the hierarchy comprises sixteen 4×4 sub-blocks (so that N=4), which sub-blocks are arrived at by splitting each 8×8 sub-block four ways. Thus, a hierarchy of levels i=0,1, and 2 is achieved with sub-blocks of size 16×16, 8×8, and 4×4, respectively. It should be noted that each of the 4×4 sub-blocks in the i=2 level of the hierarchy is childless since 4×4 is the smallest block size of interest.
In the inventive algorithm, the maximum sized block to which an orthogonal transform such as a discrete cosine transform is applied is an M×M block and the smallest block to which a discrete cosine transform is applied is N×N. Depending on the image contents of a particular M×M block, the discrete cosine transform may be applied to the M×M block as a whole, to all 4K-1 sub-blocks of size N×N which are derived from the M×M block, or to a mosaic of sub-blocks of different sizes, which sizes are given by M/2i ×M/2i where i=0,1, . . . K-1.
Thus, in the three level hierarchy, presented as an example above, the discrete cosine transform may be applied to a 16×16 block as a whole. Alternatively, depending on image contents, the discrete cosine transform may be applied individually to the sixteen 4×4 sub-blocks derived from the 16×16 block. In another case, the discrete cosine transform may be applied individually to four 4×4 brother/sister sub-blocks in the i=K-1=2 level of the hierarchy and to three 8×8 brother/sister sub-blocks in the i=K-2=1 level of the hierarchy.
To determine the blocks to which the discrete cosine transform is applied, a testing and selective merging process is utilized based on image contents.
Starting at the i=K-1 level of the hierarchy, the algorithm first checks to see if the discrete cosine transform should be applied to sub-blocks of size N×N (i.e. blocks of M/2K-1 ×M/2K-1) by using a test involving four adjacent brother/sister N×N blocks. If the test is negative, the four adjacent N×N blocks are merged to form a single (now childless) 2N×2N block (i.e. a block of size M/2K-2 ×M/2K-2). If the test is positive the discrete cosine transform is applied individually to the four brother/sister blocks of size N×N.
If the testing and merging process, when applied to the (K-1) level of the hierarchy, results in the formation of four (now childless) brother/sister 2N×2N sub-blocks in the (K-2) level of the hierarchy, the test is applied to determine if the four 2N×2N sub-blocks should be merged. If such 2N×2N sub-blocks are not merged, the discrete cosine transform is applied to the 2N×2N sub-blocks individually. Otherwise, these four 2N×2N sub-blocks are merged into a single 4N×4N (now childless) block. Similarly, if the testing and merging process, when applied to the K-1 level of the hierarchy, results in the formation of one or more 2N×2N sub-blocks in the (K-2) level of the hierarchy, but not a set of four childless brother/sister 2N×2N sub-blocks, then the discrete cosine transform is applied individually to the 2N×2N sub-blocks.
This testing and selective merger process continues up the hierarchy until the M×M block is reached in which case the discrete cosine transform is applied to the M×M block or until at some level in the hierarchy there are no possible mergers left (i.e., no sets of four childless brother/sister sub-blocks to be tested) so that the discrete cosine transform is applied to the remaining sub-blocks at that level in the hierarchy.
Illustratively, the test to determine whether or not to merge a set of brother/sister sub-blocks at a particular level of the hierarchy utilizes the mean pixel value of the sub-blocks. Let mi (k); k=1,2,3,4 represent the mean pixel value of four brother/sister sub-blocks at the ith level of the hierarchy. If there is a j≠k such that
|mi (j)-mi (k)|>ti ( 1)
do not merge; otherwise merge, where ti is a predetermined value for the ith level in the hierarchy.
Using the above-described sub-block hierarchy, and the above-described testing and selective merging steps, an M×M block is decomposed into one or more sub-blocks to which a discrete cosine transform is applied.
As indicated above, in discrete cosine transform coding, large block size introduces visible degradation if the block contains high contrast edges. Therefore, it is desirable to keep the block size as small as possible for such areas. The testing criterion based on mean pixel value described above can discriminate quite well between those areas with high contrast edges and other areas where the human visual system is not sensitive to degradation and for which larger block size should be used to take advantage of the greater compression capability offered by a larger block size.
After the block size determination procedure described above is performed, a discrete cosine transform is applied to the resulting variably sized blocks. The resulting transform coefficients in array form are compressed by thresholding and quantization in the manner discussed above. The compressed transform coefficients are then zigzag scanned to form a one dimensional sequence of transform coefficients which are then coded for transmission to a remote receiver or for storage in a memory.
In comparison to the conventional transform coding algorithm, the adaptive transform coding algorithm based on variable block size achieves the same or better visual quality with a lower average bit rate.
FIG. 1 schematically illustrates a conventional transform coding algorithm for a still image.
FIG. 2 schematically illustrates a three-level quadtree structure and its binary representation.
FIG. 3 schematically illustrates an adaptive transform coding algorithm using variable block size in accordance with an illustrative embodiment of the present invention.
A conventional transform coding algorithm for a still image is schematically illustrated by the block diagram of FIG. 1.
A still image signal to be coded comprises a two-dimensional array of pixels, e.g., 512×512 pixels. The still image signal is divided (step 12 of FIG. 1) into disjoint blocks of smaller size, e.g., 8×8 or 16×16 blocks of pixels.
A two-dimensional discrete cosine transform is applied to each block of pixels (step 14 of FIG. 1).
The two-dimensional discrete cosine transform of a two-dimensional pixel value array f(j,k) for j,k=0,1. . . n-1 can be defined as ##EQU1## for u,v=0,1, . . . N-1, where ##EQU2##
Thus, F(u,v) for u,v=0,1, . . . n-1 represents a two-dimensional array of transform coefficients.
The inverse transform is then given by ##EQU3##
The coefficients F(u,v) obtained for a block of pixels may be viewed as comprising DC and AC coefficients. The DC coefficient is F(0,0) which represents the average intensity of the block. The remainder of the coefficients (F(0,1), F(1,0)etc.) are AC coefficients.
Illustratively, the transform coefficients F(u,v) obtained by applying the discrete cosine transform to each block of pixels in the input image are normalized to a range of [-255,255] for the AC coefficients and [0,510] for the DC coefficient if the input image data range from [0,255], i.e., 8 bits per pixel.
The two-dimensional array of the transform coefficients F(u,v) obtained for each block of pixels is then compressed by thresholding and quantization (step 16 of FIG. 1).
In particular, each of the coefficients except the DC coefficient F(0,0) is thresholded and each of the coefficients is quantized.
The characteristics of the thresholding step are as follows: ##EQU4## where FT (u,v) represents the thresholded coefficients and T is a predetermined threshold value. This means that if the magnitude of a coefficient is less than T it is set to zero. In some thresholding processes, when the magnitude of a coefficient is greater than T, T is subtracted from a positive coefficient and added to a negative coefficient.
The characteristics of the quantizing step are as follows:
FTQ (u,v)=NINT(FT (u,v)/Q) (6)
where FTQ (u,v) is a thresholded and quantized transform coefficient, Q is a quantization step size and NINT is the nearest integer function. Thus, a quantized and thresholded transform coefficient FTQ (u,v) is obtained from a thresholded coefficient FT (u,v) by scaling with a step size Q and rounding to the nearest integer. The DC coefficient F(0,0) (which is not thresholded) is quantized with Q=1 to ensure intensity continuity within the image.
After compression, the two-dimensional array of transform coefficients FTQ (u,v) u=0,1, . . . n-1,v=0,1, . . . n-1 is zig-zag scanned (step 18 of FIG. 1) (see the Chen et al reference identified above) into a one-dimensional sequence of transform coefficients to facilitate encoding. The one-dimensional sequence of transform coefficients is then coded (step 20 of FIG. 1) using an entropy encoding process to form a bit stream which may be transmitted to a channel coder for transmission to a remote receiver or stored in a memory.
As indicated above, the present invention modifies the above-identified algorithm so that the pixel block size is adaptive based on image contents. Before discussing the inventive algorithm in detail it will be helpful to discuss a data structure known as a quadtree data structure. The quadtree data structure is useful for understanding the determination of block size in accordance with the inventive algorithm.
The quadtree 40 of FIG. 2 is a hierarchical data structure where each node (e.g. 42, 44) generates four children, unless it is a leaf (e.g. 46). Thus, the node 44 has four children 51, 52, 53, 54. The children 61, 62, 63, 64 of the node 46, which are shown in phantom, have been removed from the quadtree structure so that the node 46, as stated above, has no children and is a leaf. Similarly, the nodes 51, 52, 53, 54 are childless and are also leaves.
All nodes originating from the same parent are brother/sister nodes to each other. Thus, the nodes 51, 52, 53, 54 are such brother/sister nodes. In FIG. 2, non-leaf and leaf nodes are represented by logic "1" and logic "0" respectively. The node 42 at the top of the quadtree structure of FIG. 2 is a root, which is also defined as a 0th level node. If a node is an ith level node, then all the other nodes in the same generation are also ith level nodes. The children, if any, of an ith level node are (i+1)th level nodes. The quadtree structure 40 of FIG. 2 is a three level structure. The 0th level includes the root node 42. The first level includes the nodes 44, 46, 47, 48 and the second level includes the nodes 51, 52, 53, 54.
As indicated above, the present invention is an adaptive transform coding technique wherein adaptivity is achieved via a variable block size determined according to block attribute. The adaptive transform coding algorithm of the present invention is illustrated in FIG. 3.
A process (100 of FIG. 3) for choosing a block size based on image content is now considered in detail. First the image is divided into relatively large M×M blocks (step 102 of FIG. 3). Each M×M block may be viewed as a root node in a quadtree data structure.
The smallest sub-block size of interest in the algorithm is N×N where M=N×2K-1, where K is a positive integer.
Each M×M block is decomposed into 4K-1 sub-blocks of size N×N by successively splitting each block into four equally sized sub-blocks until the smallest sub-block size of N×N is reached (step 104 of FIG. 3). The sub-blocks which result from splitting an M×M block (K-1) times may be represented by a full K level quadtree structure (i.e. a K level quadtree structure in which all children are present) wherein each node represents a sub-block of pixels. In particular, an ith level node represents a sub-block of size M/2i ×M/2i.
Illustratively, a three level partition may be utilized. Thus, M may be set equal to 16 so that the largest block size of interest is 16×16 . This block corresponds to the 0th level root node 42 of the quadtree structure of FIG. 2. The 16×16 block is first divided into four 8×8 sub-blocks corresponding to the first level nodes of FIG. 2. Each 8×8 sub-block is then divided into four 4×4 sub-blocks corresponding to the second level nodes of FIG. 2. Some of the second level nodes, e.g. 61, 62, 63, 64 are shown in phantom in FIG. 2. For the reasons stated below, these nodes and the sub-blocks corresponding to them have been removed from the quadtree structure.
In accordance with the variable block size adaptive transform coding algorithm, a test is made (step 106 of FIG. 3) to determine if a discrete cosine transform should be applied to the childless sub-blocks of size N×N. The test uses a decision rule involving four adjacent childless brother/sister sub-blocks. If the test is positive the discrete cosine transform is applied (step 108 of FIG. 3) to the four childless N×N brother/sister sub-blocks. If the test is negative, these sub-blocks are merged (step 109 of FIG. 3) into one 2N×2N sub-block. The 2N×2N sub-block formed by such a merger is childless. In other words, the nodes representing the now merged N×N sub-blocks are removed from the quadtree structure so that the node representing the parent 2N×2N sub-block becomes a leaf.
In the three level example shown in FIG. 2, the test is applied to the set of brother/sister 4×4 sub-blocks represented by the nodes 51, 52, 53, 54; the set of brother/sister 4×4 sub-blocks represented by the nodes 61, 62, 63, 64; the set of brother/sister 4×4 sub-blocks represented by the nodes 81, 82, 83, 84 and the set of brother/sister 4×4 sub-blocks represented by the nodes 71, 72, 73, 74.
The 4×4 sub-blocks represented by nodes 61, 62, 63, 64 (shown in phantom) are removed by merging them into the 8×8 sub-block represented by the node 46. Thus, as a result of the test and merger steps the node 46 becomes a leaf whose binary representation, as indicated above, is a logic "0". Similarly, the 4×4 sub-blocks represented by the nodes 71, 72, 73, 74 (shown in phantom) are removed by merging them into the 8×8 sub-block represented by the node 47. On the other hand, the 4×4 sub-blocks represented by the nodes 51, 52, 53, 54 are not merged and the sub-blocks represented by the nodes 81, 82, 83, 84 are not merged. Thus, the discrete cosine transform is applied individually to these sub-blocks.
The test and selective merger process is repeated at each level of the hierarchy for each set of four sub-blocks represented by brother/sister nodes which are all leaves (i.e. which have no children and are thus represented by logic value "0") until a level in the hierarchy is reached in which there are no sets of brother/sister nodes which are all leaves or until the root node is reached. The discrete cosine transform is then applied individually to the resulting arrangement of blocks of variable size.
In the three level structure of FIG. 2, the test and selective merger process is not applied at the level i=1 because there is no set of four brother/sister nodes which are leaves. Instead, at the level i=1, there are two nodes which are leaves (46, 47) and two nodes which are not leaves (44, 48). Thus, in the example of FIG. 2, the discrete cosine transform is applied individually to the 4×4 sub-blocks represented by the nodes 51, 52, 53, 54 and the nodes 81, 82, 83, 84. The discrete cosine transform is also applied individually to the 8×8 sub-blocks represented by the nodes 46, 47.
The decision rule which determines whether to merge a particular set of four childless brother/sister sub-blocks is based on the differences in mean pixel values of the four adjacent brother/sister sub-blocks. Assuming that the blocks are ith level blocks and that the mean pixel value of the kth sub-block is mi (k), for k=1,2,3,4, then if there is a j≠k such that |mi(j)-mi(k)|>ti ; do not merge, otherwise merge.
Note, that ti is a predetermined value for the ith level in the quadtree structure. Different values may be utilized at different levels in the quadtree structure.
As indicated above, in discrete cosine transform coding, large block size introduces visible degradation if the block contains high contrast edges. Therefore, it is desirable to keep the blocksize as small as possible for these areas. The above-described process, involving use of a hierarchy of sub-blocks and testing and selective merging steps based on a decision rule using mean pixel values, serves to divide each M×M block from areas containing high contrast edges into sub-blocks of relatively small size. In contrast, M×M blocks from other image areas where the human visual system is not sensitive to degradation are divided into sub-blocks of larger size or not divided at all to achieve greater compression.
After the quadtree structure of sub-blocks is obtained, the discrete cosine transform is applied to the sub-blocks (step 108, FIG. 3). The transform coefficients resulting from application of the transform to each sub-block are then compressed by thresholding with a threshold value T and quantizing with a quantization step size Q (step 110 of FIG. 3). As indicated above, the image quality and bit rate are determined at least in part by the compression parameters T and Q. In accordance with an illustrative embodiment of the invention, different compression parameters T and Q may be chosen for sub-blocks of different size (i.e. for sub-blocks at different levels of the quadtree structure).
After compression, the transform coefficients of a sub-block of pixels are zigzag scanned (step 112 of FIG. 3) to convert the transform coefficients from an array to a sequence. The sequence of transform coefficients is then encoded (step 114 of FIG. 3) using for example entropy coding. The resulting bit stream is then sent to a channel coder for transmission to a remotely located receiver or stored in a memory. For entropy coding, different code books may be used for each block size according to its statistics.
For reconstruction of a block of pixels (i.e. an original M×M block) from the bit stream received at a remote receiver, it is necessary to inform the receiver of the quadtree structure used to partition the block. For this reason, some overhead bits are sent along with the image data. In practice, the bit value (logic "1" or logic "0") assigned to each node in the quadtree represent this overhead. This overhead is extremely trivial for a three or four level quadtree block partition structure. It should be noted that if the receiver has a priori knowledge of the number of levels in the quadtree structure, it is not necessary to send the information concerning the lowest level nodes in the quadtree structure. In the example of FIG. 2, the overhead bits may be 11010 corresponding to nodes 42, 44, 46, 48, 47, respectively. It is well worth mentioning that using different compression parameters and different code books for different sub-block sizes does not require additional overhead since the information is contained in the quadtree structure.
At a receiver, a received bit stream corresponding to a sub-block of pixels is decoded and the received transform coefficients are expanded by the appropriate quantization step size. An inverse orthogonal transform such as an inverse discrete cosine transform is then used to reconstruct the sub-block of pixels.
In an illustrative example of the inventive algorithm, the input image is 8 bits per pixel [0,255] and its size is 512×512. A three level block size partition scheme is utilized in which the largest block size of interest is 16×16 and the smallest block size of interest is 4×4. The decision thresholds ti in the block partitioning scheme are set to 10 for both i=1 and i=2 in the three level partition scheme. The partitioning scheme serves to divide the image into a mosaic of 16×16, 8×8, and 4×4 blocks depending on the contents of particular areas in the image.
After the discrete cosine transform is applied to the 16×16, 4×4 and 8×8 blocks, the transform coefficients are normalized to a range of [0,510] for the DC transform coefficients and [-255,255] for the AC transform coefficients. Note that if the discrete cosine transform is defined as in Equation (2) above, these ranges are achieved automatically when the input data ranges from [0,255].
For compression, the threshold value T is set to zero for all block sizes and the quantization step size Q is set to 1,3, and 6 for 16×16, 8×8 and 4×4 blocks, respectively.
For an image comprising text, the bit rate achieved using the variable block size adaptive transform coding algorithm is 2.2548 bits per pixel. The mean square error is 4.5 and the quadtree overhead is 0.01616 bits per pixel. In contrast, when the same "text" image is encoded using the conventional discrete cosine transform algorithm, with a 16×16 block size and a quantization step size of Q=4.0, the bit rate is 2.3639 bits per pixel and the corresponding mean square error is 66.1. Thus, the inventive algorithm provides a better quality reconstructed image at a lower bit rate than the conventional algorithm.
Finally, the above-described embodiments of the invention are intended to be illustrative only. Numerous alternative embodiments may be devised by those skilled in the art without departing from the spirit and scope of the following claims.
| Patent | Priority | Assignee | Title |
| 10015519, | Jun 07 2009 | LG Electronics Inc. | Method and apparatus for decoding a video signal |
| 10063863, | Jul 18 2003 | Microsoft Technology Licensing, LLC | DC coefficient signaling at small quantization step sizes |
| 10075731, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Video coding / decoding with re-oriented transforms and sub-block transform sizes |
| 10123038, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Video coding / decoding with sub-block transform sizes and adaptive deblock filtering |
| 10158879, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Sub-block transform coding of prediction residuals |
| 10306227, | Jun 03 2008 | Microsoft Technology Licensing, LLC | Adaptive quantization for enhancement layer video coding |
| 10390037, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Video coding/decoding with sub-block transform sizes and adaptive deblock filtering |
| 10405001, | Jun 07 2009 | LG Electronics Inc. | Method and apparatus for decoding a video signal |
| 10462494, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 10554985, | Jul 18 2003 | Microsoft Technology Licensing, LLC | DC coefficient signaling at small quantization step sizes |
| 10659793, | Jul 18 2003 | Microsoft Technology Licensing, LLC | DC coefficient signaling at small quantization step sizes |
| 10986372, | Jun 07 2009 | LG Electronics Inc. | Method and apparatus for decoding a video signal |
| 11076175, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 11659210, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 12096043, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 12120352, | Jun 07 2009 | LG Electronics Inc. | Method and apparatus for decoding a video signal |
| 5299019, | Feb 28 1992 | Samsung Electronics Co., Ltd. | Image signal band compressing system for digital video tape recorder |
| 5396567, | Nov 16 1990 | Siemens Aktiengesellschaft | Process for adaptive quantization for the purpose of data reduction in the transmission of digital images |
| 5414469, | Oct 31 1991 | International Business Machines Corporation | Motion video compression system with multiresolution features |
| 5414527, | Aug 14 1991 | Fuji Xerox Co., Ltd. | Image encoding apparatus sensitive to tone variations |
| 5416603, | Apr 30 1991 | Ricoh Company, Ltd. | Image segmentation using discrete cosine transfer data, and image data transmission apparatus and method using this image segmentation |
| 5446806, | Nov 15 1993 | National Semiconductor Corp | Quadtree-structured Walsh transform video/image coding |
| 5495538, | May 28 1993 | Xerox Corporation | Segmentation-based JPEG image artifacts reduction |
| 5596659, | Sep 01 1992 | Apple Inc | Preprocessing and postprocessing for vector quantization |
| 5649030, | Sep 01 1992 | Apple Inc | Vector quantization |
| 5717497, | Apr 30 1991 | Ricoh Company, Inc. | Image-data transmitter having function of automatic segmented image coding |
| 5729633, | Aug 30 1991 | FUJIFILM Corporation | Picture data compression coding device and method thereof |
| 5748786, | Sep 21 1994 | RICOH COMPANY, LTD , A CORPORATION OF JAPAN; Ricoh Corporation | Apparatus for compression using reversible embedded wavelets |
| 5748788, | Sep 29 1992 | Cannon Kabushiki Kaisha | Image processing method and apparatus |
| 5768434, | Nov 15 1993 | National Semiconductor Corp. | Quadtree-structured walsh transform coding |
| 5786856, | Mar 19 1996 | International Business Machines | Method for adaptive quantization by multiplication of luminance pixel blocks by a modified, frequency ordered hadamard matrix |
| 5790747, | Oct 31 1992 | Sony Corporation | Digital video signal recording apparatus with shuffling across an entire high definition image |
| 5812195, | Sep 14 1993 | Envistech, Inc. | Video compression using an iterative correction data coding method and systems |
| 5822465, | Sep 01 1992 | Apple Inc | Image encoding by vector quantization of regions of an image and codebook updates |
| 5831688, | Oct 31 1994 | Mitsubishi Denki Kabushiki Kaisha | Image coded data re-encoding apparatus |
| 5864637, | Apr 18 1995 | AMD TECHNOLOGIES HOLDINGS, INC ; GLOBALFOUNDRIES Inc | Method and apparatus for improved video decompression by selective reduction of spatial resolution |
| 5867602, | Sep 21 1994 | RICOH COMPANY, LTD , A CORPORATION OF JAPAN | Reversible wavelet transform and embedded codestream manipulation |
| 5881176, | Jun 30 1995 | RICOH CO , LTD | Compression and decompression with wavelet style and binary style including quantization by device-dependent parser |
| 5903669, | Jan 31 1995 | Canon Kabushiki Kaisha | Image processing apparatus and method |
| 5923787, | Sep 17 1993 | Ricoh Company Ltd. | Quantization device and method, inverse-quantization device and method, and image processing device and method |
| 5949422, | Jul 29 1996 | MATSUSHITA ELECTRIC INDUSTRIAL CO , LTD | Shape data compression method, shape data decompression method, shape data compression apparatus, and shape data decompression apparatus |
| 5966465, | Sep 21 1994 | RICOH CO LTD | Compression/decompression using reversible embedded wavelets |
| 5999655, | Jul 11 1996 | U S PHILIPS CORPORATION | Method and apparatus for transmission of video pictures which are segmented to form maps of picture blocks of variable block sizes |
| 5999656, | Jan 17 1997 | RICOH CO LTD , HAVING A PLACE OF BUSINESS IN TOKYO, JAPAN; Ricoh Corporation | Overlapped reversible transforms for unified lossless/lossy compression |
| 6002801, | Apr 18 1995 | Advanced Micro Devices, Inc. | Method and apparatus for improved video decompression by selection of IDCT method based on image characteristics |
| 6044172, | Dec 22 1997 | Ricoh Company, LTD | Method and apparatus for reversible color conversion |
| 6148027, | May 30 1997 | MEDIATEK, INC | Method and apparatus for performing hierarchical motion estimation using nonlinear pyramid |
| 6195465, | Sep 20 1994 | RICOH COMPANY, LTD , A CORP OF JAPAN | Method and apparatus for compression using reversible wavelet transforms and an embedded codestream |
| 6222941, | Sep 21 1994 | Ricoh Corporation | Apparatus for compression using reversible embedded wavelets |
| 6263106, | Jul 04 1995 | Hoya Corporation | Image data compression device having compression inhibiting function |
| 6314452, | Aug 31 1999 | IDX Investment Corporation | System and method for transmitting a digital image over a communication network |
| 6337929, | Sep 29 1997 | Canon Kabushiki Kaisha | Image processing apparatus and method and storing medium |
| 6339657, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Image compression of a facial region by adaptive quantization based on resolution |
| 6430713, | Jun 30 1999 | LinkedIn Corporation | System and method for constructing low complexity block coders |
| 6529634, | Nov 08 1999 | Qualcomm Incorporated | Contrast sensitive variance based adaptive block size DCT image compression |
| 6539119, | Aug 30 1993 | Sony Corporation | Picture coding apparatus and method thereof |
| 6577772, | Dec 23 1998 | LG Electronics Inc. | Pipelined discrete cosine transform apparatus |
| 6584230, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Image decoding of image content within a boundary based on a map resolution transformed signal |
| 6614938, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Image compression with orthogonal transform of blocks within a bounded shape |
| 6640013, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Image decoding using a shape map |
| 6697531, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Image coding and decoding using a position and shape map |
| 6711299, | Mar 11 1997 | UPLOAD TECHNOLOGIES, S A | Wavelet transformation of dithered quantized/reduced color pixels for color bit depth image compression and decompression |
| 6714684, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Subband image compression of blocks within a bounded shape using a resolution transform |
| 6826308, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Method and apparatus for image encoding and decoding using a shape and position map |
| 6853755, | Mar 28 2001 | Sharp Laboratories of America, Inc. | Method and apparatus for adaptive compression of scanned documents |
| 6859563, | Mar 30 2001 | RICOH CO , LTD | Method and apparatus for decoding information using late contexts |
| 6868184, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Method and apparatus for image coding and/or decoding using a position and shape map |
| 6870963, | Jun 15 2001 | Qualcomm, Inc.; Qualcomm Incorporated | Configurable pattern optimizer |
| 6873734, | Sep 21 1994 | Ricoh Corporation | Method and apparatus for compression using reversible wavelet transforms and an embedded codestream |
| 6895120, | Mar 30 2001 | RICOH CO , LTD | 5,3 wavelet filter having three high pair and low pair filter elements with two pairs of cascaded delays |
| 6898323, | Feb 15 2001 | RICOH CO LTD | Memory usage scheme for performing wavelet processing |
| 6898325, | Feb 15 2001 | Ricoh Company, Ltd. | Method and apparatus for clipping coefficient values after application of each wavelet transform |
| 6904175, | Mar 11 1997 | UPLOAD TECHNOLOGIES, S A | Image compression using an interger reversible wavelet transform with a property of precision preservation |
| 6904178, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for eliminating flicker by quantizing values based on previous quantization |
| 6925209, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for outputting a codestream as multiple tile-part outputs with packets from tiles being output in each tile-part |
| 6947874, | Nov 16 2000 | Canon Kabushiki Kaisha | Entropy coding |
| 6950558, | Mar 30 2001 | RICOH CO , LTD , A CORPORATION OF JAPAN | Method and apparatus for block sequential processing |
| 6973217, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for sending additional sideband information in a codestream |
| 6975773, | Jul 30 2002 | Qualcomm Incorporated; QUALCOMM INCORPORATED A DELAWARE CORPORATION | Parameter selection in data compression and decompression |
| 6983075, | Feb 15 2001 | Ricoh Co., Ltd | Method and apparatus for performing selective quantization by manipulation of refinement bits |
| 6990247, | Sep 20 1994 | Ricoh Corporation | Multiple coder technique |
| 6996283, | Nov 16 2001 | Qualcomm Incorporated | Block size assignment using local contrast ratio |
| 7003168, | Mar 11 1997 | UPLOAD TECHNOLOGIES, S A | Image compression and decompression based on an integer wavelet transform using a lifting scheme and a correction method |
| 7006697, | Mar 30 2001 | RICOH CO , LTD | Parallel block MQ arithmetic image compression of wavelet transform coefficients |
| 7016545, | Sep 21 1994 | RICOH CORPORATION, LTD A CORPORATION OF JAPAN | Reversible embedded wavelet system implementation |
| 7024046, | Apr 18 2000 | IDX Investment Corporation | System and method for the lossless progressive streaming of images over a communication network |
| 7054493, | Sep 21 1994 | Ricoh Co., Ltd. | Context generation |
| 7054495, | Nov 10 1994 | Kabushiki Kaisha Toshiba | Image data coding and/or decoding system capable of high-efficient coding |
| 7062101, | Mar 30 2001 | RICOH CO , LTD | Method and apparatus for storing bitplanes of coefficients in a reduced size memory |
| 7062103, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for specifying quantization based upon the human visual system |
| 7068849, | Sep 21 1994 | Ricoh Corporation | Method and apparatus for compression using reversible wavelet transforms and an embedded codestream |
| 7072520, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for selecting layers for quantization based on sideband information |
| 7076104, | Sep 21 1994 | Ricoh Corporation | Compression and decompression with wavelet style and binary style including quantization by device-dependent parser |
| 7079690, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for editing an image while maintaining codestream size |
| 7088858, | Dec 17 1999 | Canon Kabushiki Kaisha | Digital signal coding with division into tiles |
| 7088869, | Mar 30 2001 | Ricoh Co., Ltd. | 5,3 wavelet filter having three high pair and low pair filter elements with two pairs of cascaded delays |
| 7095900, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for performing scalar quantization with a power of two step size |
| 7095907, | Jan 10 2002 | RICOH CO LTD A CORPRORATION OF JAPAN | Content and display device dependent creation of smaller representation of images |
| 7120305, | Apr 16 2002 | RICOH CO , LTD | Adaptive nonlinear image enlargement using wavelet transform coefficients |
| 7139434, | Sep 21 1994 | Ricoh Co., Ltd. | Decoding with storage of less bits for less important data |
| 7164804, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for eliminating flicker by quantizing values based on previous quantization |
| 7167589, | Sep 21 1994 | Ricoh Co., Ltd. | Disk read technique |
| 7167592, | Sep 21 1994 | Ricoh Co., Ltd. | Method and apparatus for compression using reversible wavelet transforms and an embedded codestream |
| 7215820, | Sep 21 1994 | Ricoh Co., Ltd. | Method and apparatus for compression using reversible wavelet transforms and an embedded codestream |
| 7221761, | Sep 18 2000 | Sharp Kabushiki Kaisha | Error resilient digital video scrambling |
| 7227999, | Sep 21 1994 | Ricoh Co., Ltd. | Printing system application using J2K |
| 7280252, | Dec 19 2001 | RICOH CO , LTD | Error diffusion of multiresolutional representations |
| 7289677, | Sep 21 1994 | Ricoh Co., Ltd. | Reversible embedded wavelet system implementation |
| 7298912, | Mar 30 2001 | Ricoh Co., Ltd. | Method and apparatus for assigning codeblocks to coders operating in parallel |
| 7315648, | Dec 17 1999 | Canon Kabushiki Kaisha | Digital signal coding with division into tiles |
| 7321695, | Sep 21 1994 | Ricoh Co., Ltd. | Encoder rate control |
| 7376279, | Dec 14 2000 | IDX Investment Corporation | Three-dimensional image streaming system and method for medical images |
| 7382968, | Feb 21 2002 | Sovereign Peak Ventures, LLC | Recording apparatus and computer-readable program |
| 7388993, | Jul 30 2002 | Qualcomm Incorporated | Parameter selection in data compression and decompression |
| 7397963, | Mar 30 2001 | Ricoh Co., Ltd. | Method and apparatus for storing bitplanes of coefficients in a reduced size memory |
| 7418142, | Sep 20 1994 | Ricoh Company, Ltd. | Method for compression using reversible embedded wavelets |
| 7454074, | Apr 18 2000 | General Electric Company | System and method for the lossless progressive streaming of images over a communication network |
| 7457473, | Mar 30 2001 | Ricoh Co., Ltd. | Method for block sequential processing |
| 7474791, | Jan 10 2002 | Ricoh Co., Ltd. | Content and display device dependent creation of smaller representations of images |
| 7477792, | Feb 15 2001 | Ricoh Co., Ltd. | Method and apparatus for performing progressive order conversion |
| 7580584, | Jul 18 2003 | Microsoft Technology Licensing, LLC | Adaptive multiple quantization |
| 7581027, | Jun 27 2001 | RICOH CO , LTD | JPEG 2000 for efficent imaging in a client/server environment |
| 7593582, | Jul 30 2002 | Qualcomm Incorporated | Systems and methods using parameter selection in data compression and decompression |
| 7602851, | Jul 18 2003 | Microsoft Technology Licensing, LLC | Intelligent differential quantization of video coding |
| 7630563, | Jul 19 2001 | Qualcomm Incorporated | System and method for decoding digital image and audio data in a lossless manner |
| 7634145, | Sep 21 1994 | Ricoh Corporation | Compression and decompression with wavelet style and binary style including quantization by device-dependent parser |
| 7738554, | Jul 18 2003 | Microsoft Technology Licensing, LLC | DC coefficient signaling at small quantization step sizes |
| 7801383, | May 15 2004 | Microsoft Technology Licensing, LLC | Embedded scalar quantizers with arbitrary dead-zone ratios |
| 7813564, | Mar 30 2007 | Monument Peak Ventures, LLC | Method for controlling the amount of compressed data |
| 7974340, | Apr 07 2006 | Microsoft Technology Licensing, LLC | Adaptive B-picture quantization control |
| 7995649, | Apr 07 2006 | ZHIGU HOLDINGS LIMITED | Quantization adjustment based on texture level |
| 8031955, | Apr 03 2006 | Fuji Xerox Co., Ltd. | Image processing apparatus, image processing method, medium storing program, and computer data signal |
| 8059721, | Apr 07 2006 | Microsoft Technology Licensing, LLC | Estimating sample-domain distortion in the transform domain with rounding compensation |
| 8130828, | Apr 07 2006 | Microsoft Technology Licensing, LLC | Adjusting quantization to preserve non-zero AC coefficients |
| 8131095, | Feb 03 2003 | NINTENDO EUROPEAN RESEARCH AND DEVELOPMENT | Process and device for the compression of portions of images |
| 8184694, | May 05 2006 | Microsoft Technology Licensing, LLC | Harmonic quantizer scale |
| 8189933, | Mar 31 2008 | Microsoft Technology Licensing, LLC | Classifying and controlling encoding quality for textured, dark smooth and smooth video content |
| 8218624, | Jul 18 2003 | Microsoft Technology Licensing, LLC | Fractional quantization step sizes for high bit rates |
| 8224105, | Oct 27 2008 | PANASONIC SEMICONDUCTOR SOLUTIONS CO , LTD | Method for coding image, image coding device and image pickup system |
| 8238424, | Feb 09 2007 | Microsoft Technology Licensing, LLC | Complexity-based adaptive preprocessing for multiple-pass video compression |
| 8243797, | Mar 30 2007 | Microsoft Technology Licensing, LLC | Regions of interest for quality adjustments |
| 8244033, | May 13 2008 | Canon Kabushiki Kaisha | Image encoding apparatus, image decoding apparatus, and control method thereof |
| 8249145, | Apr 27 2006 | Microsoft Technology Licensing, LLC | Estimating sample-domain distortion in the transform domain with rounding compensation |
| 8331438, | Jun 05 2007 | Microsoft Technology Licensing, LLC | Adaptive selection of picture-level quantization parameters for predicted video pictures |
| 8358864, | Jun 11 2009 | Sony Corporation | Image processing apparatus and image processing method |
| 8422546, | May 25 2005 | Microsoft Technology Licensing, LLC | Adaptive video encoding using a perceptual model |
| 8442337, | Apr 18 2007 | Microsoft Technology Licensing, LLC | Encoding adjustments for animation content |
| 8472741, | Jun 11 2009 | Sony Corporation | Image processing apparatus and image processing method |
| 8498335, | Mar 26 2007 | Microsoft Technology Licensing, LLC | Adaptive deadzone size adjustment in quantization |
| 8503536, | Apr 07 2006 | Microsoft Technology Licensing, LLC | Quantization adjustments for DC shift artifacts |
| 8565298, | Sep 21 1994 | Ricoh Co., Ltd. | Encoder rate control |
| 8576908, | Mar 30 2007 | Microsoft Technology Licensing, LLC | Regions of interest for quality adjustments |
| 8588298, | May 05 2006 | Microsoft Technology Licensing, LLC | Harmonic quantizer scale |
| 8638864, | Jul 16 1997 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
| 8711925, | May 05 2006 | Microsoft Technology Licensing, LLC | Flexible quantization |
| 8743949, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Video coding / decoding with re-oriented transforms and sub-block transform sizes |
| 8767822, | Apr 07 2006 | ZHIGU HOLDINGS LIMITED | Quantization adjustment based on texture level |
| 8798159, | Oct 23 2009 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video and method and apparatus for decoding video, based on hierarchical structure of coding unit |
| 8817868, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Sub-block transform coding of prediction residuals |
| 8873643, | Jul 16 1997 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
| 8891618, | Oct 23 2009 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video and method and apparatus for decoding video, based on hierarchical structure of coding unit |
| 8891631, | Oct 23 2009 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video and method and apparatus for decoding video, based on hierarchical structure of coding unit |
| 8891632, | Oct 23 2009 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video and method and apparatus for decoding video, based on hierarchical structure of coding unit |
| 8897359, | Jun 03 2008 | Microsoft Technology Licensing, LLC | Adaptive quantization for enhancement layer video coding |
| 8897369, | Oct 23 2009 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video and method and apparatus for decoding video, based on hierarchical structure of coding unit |
| 8908768, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Video coding / decoding with motion resolution switching and sub-block transform sizes |
| 8942296, | Jul 16 1997 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
| 8989274, | Oct 23 2009 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video and method and apparatus for decoding video, based on hierarchical structure of coding unit |
| 9060163, | Jul 16 1997 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
| 9060181, | Jul 16 1997 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
| 9077959, | Jul 16 1997 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
| 9185418, | Jun 03 2008 | Microsoft Technology Licensing, LLC | Adaptive quantization for enhancement layer video coding |
| 9258570, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Video coding / decoding with re-oriented transforms and sub-block transform sizes |
| 9264705, | Jul 16 1997 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
| 9264714, | Jun 11 2009 | Sony Corporation | Image processing apparatus and image processing method |
| 9300972, | Feb 09 2009 | SAMSUNG ELECTRONICS CO , LTD | Video encoding method and apparatus using low-complexity frequency transformation, and video decoding method and apparatus |
| 9313509, | Jul 18 2003 | Microsoft Technology Licensing, LLC | DC coefficient signaling at small quantization step sizes |
| 9414055, | Oct 23 2009 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding video and method and apparatus for decoding video, based on hierarchical structure of coding unit |
| 9414089, | Jun 11 2009 | Sony Corporation | Image processing apparatus and image processing method |
| 9432686, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Video coding / decoding with motion resolution switching and sub-block transform sizes |
| 9456216, | Dec 17 2001 | Microsoft Technology Licensing, LLC | Sub-block transform coding of prediction residuals |
| 9485512, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 9571840, | Jun 03 2008 | Microsoft Technology Licensing, LLC | Adaptive quantization for enhancement layer video coding |
| 9635368, | Jun 07 2009 | LG Electronics Inc | Method and apparatus for decoding a video signal |
| 9838719, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 9838720, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 9838721, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 9838722, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 9888259, | Feb 23 2009 | Korea Advanced Institute of Science and Technology; Korean Broadcasting System | Video encoding method for encoding division block, video decoding method for decoding division block, and recording medium for implementing the same |
| 9967561, | May 05 2006 | Microsoft Technology Licensing, LLC | Flexible quantization |
| Patent | Priority | Assignee | Title |
| 4261018, | Jun 18 1979 | Bell Telephone Laboratories, Incorporated | Progressive image transmission |
| 4831659, | Oct 28 1985 | Hitachi, Ltd. | Method for image data coding |
| 4944023, | May 19 1987 | Ricoh Company, Ltd. | Method of describing image information |
| 4947447, | Apr 24 1986 | Hitachi, Ltd. | Method for data coding |
| 5021891, | Feb 27 1990 | QUALCOMM INCORPORATED A CORPORATION OF DELAWARE | Adaptive block size image compression method and system |
| 5060285, | May 19 1989 | Verizon Laboratories Inc | Hierarchical variable block size address-vector quantization using inter-block correlation |
| Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
| Mar 23 1992 | Bell Communications Research, Inc. | (assignment on the face of the patent) | / | |||
| Mar 16 1999 | BELL COMMUNICATIONS RESEARCH, INC | Telcordia Technologies, Inc | CHANGE OF NAME SEE DOCUMENT FOR DETAILS | 010263 | /0311 | |
| Mar 15 2005 | Telcordia Technologies, Inc | JPMORGAN CHASE BANK, N A , AS ADMINISTRATIVE AGENT | SECURITY AGREEMENT | 015886 | /0001 | |
| Jun 29 2007 | JPMORGAN CHASE BANK, N A , AS ADMINISTRATIVE AGENT | Telcordia Technologies, Inc | TERMINATION AND RELEASE OF SECURITY INTEREST IN PATENT RIGHTS | 019520 | /0174 | |
| Jun 29 2007 | Telcordia Technologies, Inc | WILMINGTON TRUST COMPANY, AS COLLATERAL AGENT | SECURITY AGREEMENT | 019562 | /0309 | |
| Feb 20 2009 | Wilmington Trust Company | Telcordia Technologies, Inc | RELEASE OF SECURITY INTEREST | 022408 | /0410 | |
| Mar 13 2009 | CHEN, CHENG-TIE | BELL COMMUNICATIONS, RESEARCH, INC | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 022390 | /0907 | |
| Jun 16 2009 | Telcordia Technologies, Inc | Telcordia Licensing Company LLC | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 022878 | /0821 | |
| Jan 28 2010 | Telcordia Licensing Company LLC | TTI Inventions C LLC | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 025331 | /0674 | |
| Apr 30 2010 | WILMINGTON TRUST COMPANY, AS COLLATERAL AGENT | Telcordia Technologies, Inc | RELEASE | 024515 | /0622 |
| Date | Maintenance Fee Events |
| Dec 08 1995 | ASPN: Payor Number Assigned. |
| Jan 08 1997 | M183: Payment of Maintenance Fee, 4th Year, Large Entity. |
| Feb 14 2001 | M184: Payment of Maintenance Fee, 8th Year, Large Entity. |
| Feb 22 2005 | M1553: Payment of Maintenance Fee, 12th Year, Large Entity. |
| Nov 05 2009 | ASPN: Payor Number Assigned. |
| Nov 05 2009 | RMPN: Payer Number De-assigned. |
| Date | Maintenance Schedule |
| Aug 31 1996 | 4 years fee payment window open |
| Mar 03 1997 | 6 months grace period start (w surcharge) |
| Aug 31 1997 | patent expiry (for year 4) |
| Aug 31 1999 | 2 years to revive unintentionally abandoned end. (for year 4) |
| Aug 31 2000 | 8 years fee payment window open |
| Mar 03 2001 | 6 months grace period start (w surcharge) |
| Aug 31 2001 | patent expiry (for year 8) |
| Aug 31 2003 | 2 years to revive unintentionally abandoned end. (for year 8) |
| Aug 31 2004 | 12 years fee payment window open |
| Mar 03 2005 | 6 months grace period start (w surcharge) |
| Aug 31 2005 | patent expiry (for year 12) |
| Aug 31 2007 | 2 years to revive unintentionally abandoned end. (for year 12) |